For a faster way to BAN cheaters

[tempest]

- Detect client version if not updated server tells 'em to update <-- like what valve does

Masterserver does that now afaik.

[V-Man]

Now you're talking! Let's see here...

1) Customizable ban time is a great idea. I'd like to see that.
2) Kick time 20 minutes... Nah, that's the purpose of a ban. Kicks are meant to merely disconnect the person (usually as a warning before being banned).
3) DNS analysis: Can you expand on this concept, like exactly what about the DNS is analyzed?
4) Client version detection is currently implemented, but easy to spoof. If your client has a specific version (and therefore a corresponding protocol), it can only connect to servers on that same protocol.
5) There's nothing wrong with where edit mode is right now.

[vonunov]

I assume DNS means using the full DNS host mask like you might see on IRC 254.254.254.254-123456.prtlnd.or.cockscable.net or something

[eynstyne]

The DNS concept for example would like what Jack said:
254.254.254.254-123456.prtlnd.or.cockscable.net

Gives a more human readable idea of where the target is coming from rather than a bunch of numbers.
Port scanning open/closed/filtered ports manually with a tool like nmap by admins would better identify the target, but not alot of ISPs look too kindly on it so do not overuse it.
e.g : If a known target has a publicly open SSH, FTP, RDP and OS fingerprint ids as Windows 2003 server
and the suspect cheater uses a different ip should still yield the same ports & OS.

If a suspect cheater is using a whole bunch of IPs then the admin has the option to go with their gut and do a ban / blacklist <-- but NOT blacklist the isp DNS since other ppl can have the same provider and will falsely blacklist them.

I suggested edit mode to be moved because it is too simple to enable in multiplayer without coop. But realizing that coop mode does serve a purpose, to dynamically craft a map, ignore this idea.

Customizable ban time is perfect for admins because they can decide if an offense truly deserves the time given. (Does Crime fit Time?)

I forgot about the 1.04 and how the protocol doesn't work with newer ones so that's fine. But comparing what steam did is obliterate the old protocol completely. Using 1 tool called steamless could accomplish mediocre gameplay with bots and practically 4 players only on an old protocol. Not exactly high paced gameplay. :/

[tempest]

http://whois.domaintools.com/1.2.3.4

And all that without changing a single line of code ;)

Port scanning open/closed/filtered ports manually with a tool like nmap by admins would better identify the target, but not alot of ISPs look too kindly on it so do not overuse it.

Don't you dare. Btw, you might automatically land on quite a few iptables blacklists if you do that. No thanks.

I suggested edit mode to be moved because it is too simple to enable in multiplayer without coop. But realizing that coop mode does serve a purpose, to dynamically craft a map, ignore this idea.

And what would keep cheaters from switching of gravity, anyway?

I forgot about the 1.04 and how the protocol doesn't work with newer ones so that's fine. But comparing what steam did is obliterate the old protocol completely. Using 1 tool called steamless could accomplish mediocre gameplay with bots and practically 4 players only on an old protocol. Not exactly high paced gameplay. :/

Er, what?

[eynstyne]

http://whois.domaintools.com/1.2.3.4

And all that without changing a single line of code ;)

(Today 01:57PM)eynstyne Wrote: Port scanning open/closed/filtered ports manually with a tool like nmap by admins would better identify the target, but not alot of ISPs look too kindly on it so do not overuse it.

Don't you dare. Btw, you might automatically land on quite a few iptables blacklists if you do that. No thanks.

(Today 01:57PM)eynstyne Wrote: I suggested edit mode to be moved because it is too simple to enable in multiplayer without coop. But realizing that coop mode does serve a purpose, to dynamically craft a map, ignore this idea.

And what would keep cheaters from switching of gravity, anyway?

(Today 01:57PM)eynstyne Wrote: I forgot about the 1.04 and how the protocol doesn't work with newer ones so that's fine. But comparing what steam did is obliterate the old protocol completely. Using 1 tool called steamless could accomplish mediocre gameplay with bots and practically 4 players only on an old protocol. Not exactly high paced gameplay. :/

Er, what?

ok... do you have any ideas then?

[tempest]

I never have ideas, but I'm quite good at telling others why their ideas don't make sense.

EDIT: Oh, wait, I do have ideas.
1) Ban votes can't be disabled.
2) A vote succeeds if more than 50% of the given votes are positive, not if more than 50% of the players on the server voted yes.
I guess these two would solve 90% of the cheating problems.

[Luc@s]

I never have ideas, but I'm quite good at telling others why their ideas don't make sense.

EDIT: Oh, wait, I do have ideas.
1) Ban votes can't be disabled.
2) A vote succeeds if more than 50% of the given votes are positive, not if more than 50% of the players on the server voted yes.
I guess these two would solve 90% of the cheating problems.

so if I call a vote there is 1 yes on 1 vote so 1/1 => 100% of yes => vote passes directly.

[tempest]

Well, it would have to wait until the time runs out and not check it continously. Currently, a vote takes max. 20 seconds afaik. Of course, if an admin votes or if more than half of the people on the server have voted yes (or no), it would still end immediately. What I mean is just that votes shouldn't fail because half or less of the players voted.

[V-Man]

Aye, too many people don't care about voting.
Either that, or too many people hit F1 without even thinking about it.

[Zomg!]

A better soultion instead of all of these half assed ones would be a open source anti cheat developed from the ground up. lol @ this game now

[noerrorsfound]

I believe an open source anticheat would easily be bypassed. They can look at the code and see what checks are made and how to get around them, or edit everything out and compile their own useless version of the anticheat.

[Zomg!]

They could do something where every client analyzes one another and reports to the masterserver when multiple clients detect fishy behavior from one client. So that even if you know what clients check for you would have to have the majority of clients running the hacked anti cheat to get away with it. Which unless orgnaized would stop small cheaters with major obvious things.

[Ronald_Reagan]

I believe with that the bandwidth required would increase substantially (not a network techie) but I think that would be a problem with it.

[tempest]

There is no technical solution to prevent people from cheating. Period.
Even the big AAA titles secured by million-dollar anticheat software have serious problems with it. Let alone an open-source game.